Seedset protoclass

[sjanssen2]

Adressing issue #16 : adding seedset functionality to last outstanding prototype function 'protoclass'

[coveralls]

Coverage remained the same at 100.0% when pulling 57a45a0 on sjanssen2:seedset_protoclass into ad1a207 on biocore:master.

[sjanssen2]

@qiyunzhu ready for a review

[qiyunzhu]

Good job! @sjanssen2 The rationale is very clear. So the order of seeds is not important. The program just picks random one each time until all seeds are gone. Is that right?

[tkosciol]

I cannot say that I understand this PR fully, but the code looks good!
@qiyunzhu I'm not able to merge, because I'm not a member of this project.

[qiyunzhu]

@tkosciol Thank you! The code is to let the prototype selection function take a pre-defined set of "seeds", which are guaranteed to be selected as prototypes. The remaining prototypes will be calculated based on them.

[sjanssen2]

Hi @qiyunzhu I never replied to your question. You are right: elements of the seedset are made prototypes in a random ordering until all are "consumed".
Since at least |seedset| elements will be returned as prototypes (even if the user set num_prototypes to a smaller number) the ordering should not matter.